Motivation and Hypothesis

College-age students and young adults are commonly thought of as out-of-touch with the world around them and have a reputation of being self-absorbed rather than politically aware. However, being on a campus surrounded by young adults in which there is a strong political atmosphere, there are indications all around that students must be engaging with the news. I hope to find out how, and how often students are engaging with news, and whether gender has a significant effect on the frequency or pathway used to gain insight on current events. According to How Students Engage With News by Project Information Literacy (PIL), a study was conducted that surveyed 5,844 students at 11 U.S. colleges, universities, and community colleges to understand the practices of young news consumers. The main takeaways from this study concluded that most college students get their news from face-to-face discussions and social media, primarily– and specifically via Facebook. I am curious to find out if this trend holds true among Emory students. In another study done by Associated Press, it was found that 60% of students don’t necessarily seek out political news but instead, “bump” into it on social media, whereas the other 40% do actively seek out this information. While I didn’t ask this question specifically, I hope I can infer about how many Emory Economics 220 students seek out news, and how many “bump” into it.

The Data Set

After surveying students in all sections of Economics 220 and removing any instances of NAs, the sample I used consisted of 44 females and 67 males, for a total of 111 observations. The questions I used to find out which gender is more in-touch with current events and how they got their news were, respectively, “On average, how many times a week do you read or watch a reliable news source?” and “How do you consume news or current events: newspapaer, social media, radio/podcasts, or conversation?” in conjunction with the gender of the student. However, it is important to remember with this dataset that this group of surveyed students may not be representative of all Emory students’ news-reading habits, as this group is evidently interested in economics which is often a hot topic in the news. This could overestimate and misrepresent the data for the entire student body population.

# Choose the particular variables interested in:
# sex, q108, and q201
prodata <- Prodata %>% select(sex, q108, q201)
# Remove any instances containing 'na'
PD <- prodata %>% filter(!is.na(sex), !is.na(q108), 
    !is.na(q201))

# Making q201 a factor variable, order each
# category
PD$q201 <- factor(PD$q201, order = TRUE, levels = c("conversation", 
    "social media", "radio/podcasts", "newspaper"), 
    labels = c("conversation", "social media", "radio/podcasts", 
        "newspaper"))
# check the variables
summary(PD)

News Consumption by Gender

The table below shows the average number of news sources read or watched, the most frequently used pathway to get news, the maximum number of news souces read or watched, and the minimum number of sources read or watched, and the median number read or watched for each gender. According to this table, males read or watch more news sources on average than females. Both males and females appear to get most of their news information via social media networks. According to the study done by PIL, the most common pathway for students to engage with news is conversation, but this data challenges that notion.

# make sex var a factor and create a function to
# reveal frequency
PD$sex <- factor(PD$sex)
myfun <- function(x) {
    names(table(x))[table(x) == max(table(x))]
}
# sort data wanted so it is easy to create a table
Pd <- PD %>% group_by(sex) %>% summarise(`average sources read` = mean(q108), 
    `median sources read` = median(q108), `most common way of consuming news` = myfun(q201), 
    `maximum read` = max(q108), `minimum read` = min(q108))
# create a nice-looking table of the summary
# statistics above
kable(Pd, type = "html") %>% kable_styling(bootstrap_options = "responsive", 
    full_width = T)
sex average sources read median sources read most common way of consuming news maximum read minimum read
Female 3.681818 3.0 social media 20 0
Male 5.102941 3.5 social media 25 0

As we can see from this table, the means for each gender are greatly different, but the median is only slightly larger for males than for females. In addition to this, we see larger means for both genders than the respective medians. These two pieces of data indicate there must be outliers reading far more news articles than most of their peers.

# graph distribution of number of articles read by
# gender
PD_sex <- PD %>% group_by(sex)
p2 <- ggplot(data = PD_sex, mapping = aes(x = sex, 
    y = q108, fill = sex)) + geom_violin(alpha = 0.7) + 
    ylab("Number of Sources Read") + xlab("Sex") + 
    ggtitle("Distribution of News Consumption by Gender - Violin Plot") + 
    theme(plot.title = element_text(hjust = 0.5))
ggplotly(p2)
# Here we can see a larger number of females
# reading between 0-5 articles compared to males.
# There is evidently a more extensive distribution
# of news-reading among the males than the females.

Here, we can see a larger number of females reading between 0-5 articles compared to males. There is evidently a more extensive distribution of news-reading among the males than the females. There is a clear outlier in the male group who reads or watches, on average, 25 articles per week. Both females and males have the largest density around the same number of articles, between 2 and 3. I would be interested to do the same survey with students of other classes as well as faculty members on Emory’s campus to find if working adults tend to read or watch more news sources on average than students of varying majors.

Most Informed and Least Informed Observations

The following table shows the male and female in Economics 220 who read or watch the most news articles and their most common form of news engagement. It is evident these students are the outliers mentioned in the table above, which is why I chose to include the median in addition to the mean.

# most informed female, find the observation with
# the max number of articles
most_informed_female <- PD %>% filter(sex == "Female") %>% 
    arrange(desc(q108))
smartfm <- head(most_informed_female, n = 1)
# most informed male, fine the male obs with the
# max number of variables
most_informed_male <- PD %>% filter(sex == "Male") %>% 
    arrange(desc(q108))
smartm <- head(most_informed_male, n = 1)
# combine the two obs to view in a kable ot see
# where these outliers will fall
smart <- rbind(smartm, smartfm)
kable(smart, col.names = c("sex", "number of sources", 
    "media source")) %>% kable_styling(bootstrap_options = "responsive", 
    full_width = T)
sex number of sources media source
Male 25 newspaper
Female 20 social media

Media Distribution Among Students

# group data set by media pathway
PD_media <- PD %>% group_by(q201)

# plot distribution of each media pathway
PD_media <- PD %>% group_by(q201)
p4 <- ggplot(data = PD_media, mapping = aes(x = q201, 
    y = q108, fill = q201)) + geom_violin(alpha = 0.7) + 
    ylab("Number of Sources Read") + xlab("Media Type") + 
    ggtitle("News Consumption Distribution by Media Type - Violin Plot") + 
    theme(plot.title = element_text(hjust = 0.5))
ggplotly(p4)

This plot shows the density distribution of the use of each media pathway. It is interesting to see a fatter distribution for conversation among the students reading 1 or 2 articles per week. Social media users are clustering around 2.5 articles per week with the exception of a few students reading 15-20 articles per week. Radio and podcast listeners are clumped around 5 articles per week with a max of 14. It appears newspaper readers are very stretched out between 0 and 25 articles per week. Below we can analyze the percent breakdown of the use of each media pathway by students.

# Finding the number of instances of each news
# source to put into a vector for the pie chart
media_convo <- PD %>% filter(q201 == "conversation")  # 10 users
media_sm <- PD %>% filter(q201 == "social media")  #62 users
media_aud <- PD %>% filter(q201 == "radio/podcasts")  #12 users
media_news <- PD %>% filter(q201 == "newspaper")  #27 users

# 3D Exploded Pie Chart to show percent break
# down/use of each media type

slices <- c(10, 62, 12, 27)
colors <- c("sienna2", "khaki1", "turquoise", "lightsteelblue")
pct <- round(slices/sum(slices) * 100)
lbls <- paste(pct)  # add percents to labels
lbls <- paste(lbls, "%", sep = "")  # add % to labels
pie3D(slices, labels = lbls, explode = 0.1, main = "How Emory College Students Engage With News", 
    col = colors)
legend("topright", c("conversation", "social media", 
    "radio/podcasts", "newspaper"), fill = colors, 
    cex = 0.7)

This pie chart clearly shows an overwhelming use of social media to receive news information, followed by newspapers, radio/podcasts, and lastly conversation. This data follows the trends found in the Associated Press (AP) article and counters the PIL study, which says conversation should have the largest slice. As mentioned before, the AP article says about 60% of students “bump” into social media whereas 40% actively seek it out. This could be very close to true among Emory students as well, as nearly 60% (56%) of students claim to get their news through social media, networks in which they are probably doing something else and happen to run into news articles and links. 35% of students use newspapers and radio/podcasts which are more likely to be actively searched for, however, I cannot say this for certain and would have to complete another survey to find out.

Media Distribution by Gender

To breakdown media use by gender in order to compare how males and females receive their news we can refer to the box plots and piecharts below.

These boxplots further emphasize that both females and males tend to utilize social media the most, which we will also see in the following pie charts. These plots help us to visualize how many articles and what kind of news platform they are using. We see a much more extensive trend among males than females in all modes of finding news other than social media.

These pie charts simply split the first pie chart by gender so as to show how female and male Emory students differ in getting news. As expected from the first pie chart, both males and females are overwhelmingly using social media. The stark difference is between the use of newspapers and radio/podcasts, as males are using far more newspapers and females seem to use far more podcasts and radio sources. This would make for an interesting study to see if men and women differ in auditory and visual learning styles.

Conclusion

In conclusion, it appears males may tend to read more news articles than females, but that both genders primarily get their news through social media. To find out whether or not male Emory students actually read or watch a different amount of articles than female students, I will perform a two-sided t-test comparing the means to see if the mean for males is significantly different than that of the females.

\(H_{0}\): Number of mean articles read by males \(=\) Number of mean articles read by females

\(H_{1}\): Number of mean articles read by males \(\neq\) Number of mean articles read by females

t.test(q108 ~ sex, data = PD)  #fail to reject the null at the 0.05 significance level because P value is greater than 0.05. This indicates the number of sources males read is not significantly greater than the amount females read. 
## 
##  Welch Two Sample t-test
## 
## data:  q108 by sex
## t = -1.6931, df = 107.34, p-value = 0.09334
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -3.084992  0.242746
## sample estimates:
## mean in group Female   mean in group Male 
##             3.681818             5.102941

The t-test reports a p-value (0.09334) that is slightly bigger than the significance level of 0.05. In this case, we fail to reject the null hypothesis that male and females read the same number of articles per week on average, as the difference is not statistically significant at a 95% confidence level giving a confidence interval of [-3.084992, 0.242746]. This result is cohesive with the result found in a similar study done by the College Media Review, which found male respondents read more than female respondents.

References

Cohen, D. and Head, A. and MacMillan, M. and Metaxes, P. and Wihbey, J. How Students Engage With News, Project Information Literacy, 2018. News Report, https://www.projectinfolit.org/uploads/2/7/5/4/27541717/newsreport.pdf

Hedrick, Jeffrey. Research (Vol.52) Do college students wants to see political news in their newspaper?, The College Media Review: Journal of College Media Association

Survey: Young adults do consume news, in their own way, Associated Press, 2015, Online Article, https://www.ap.org/ap-in-the-news/2015/survey-young-adults-do-consume-news-in-their-own-way